Photo of printed text with numbers indicating the amount of relationships

Konzeptionelle Dichte

NLP C++ ObjC Moodboard Computational Linguistics Machine Learning Human-Computer Interaction UdK 2015

'Konzeptionelle Dichte' explores natural language processing (NLP) and how computers can recognize semantic relationships between words. Using a prototype system for automated image retrieval, I examine how machines form human-like associations and the challenges posed by ambiguity and semantic relations. The project combines linguistics and computer science, offering insights into computational linguistics, machine learning and cognitive models of language processing.

Photo of printed text with list of weighted relationships between two words Photo of a text from the hitchhikers guide to the galaxy Three different meanings for the word poster Examples for relationships between words A transparent layer with black markings above a printed text A transparent layer with black markings above a printed text A grid of pictures representing a moodboard Definition of full synsets A visual hierachy of hypernymy and hyponymy A list of lexical top levels Lexical top levels, frequency count and word sense disambiguation Gloss overlaps Visual representation of a moodboard software algorithm Visual representation of a moodboard software algorithm Example for word disambiguation Example for word disambiguation and relative weights Example for bigrams Natural language processing for the german word Danke